Skip site navigation (1)Skip section navigation (2)
Date:      Tue, 7 Feb 2012 13:40:01 -0500
From:      John Baldwin <jhb@freebsd.org>
To:        freebsd-current@freebsd.org, davidxu@freebsd.org
Subject:   Re: [ptrace] please review follow fork/exec changes
Message-ID:  <201202071340.01126.jhb@freebsd.org>
In-Reply-To: <4F30968B.5060101@gmail.com>
References:  <749E238A-A85F-4264-9DEB-BCE1BBD21C9D@juniper.net> <4F2094B4.70707@juniper.net> <4F30968B.5060101@gmail.com>

next in thread | previous in thread | raw e-mail | index | archive | help
On Monday, February 06, 2012 10:12:11 pm David Xu wrote:
> On 2012/1/26 7:48, Dmitry Mikulin wrote:
> > <snip>
> 
> > The debugger needs to intercept fork() in both parent and child so it 
> > can detach from the old process and attach to the new one. Maybe it'll 
> > make more sense in the context of gdb changes. Should I send them too? 
> > Don't think Marcel included that patch...
> >
> >>
> >> Does the orphan list change intended to not lost the child after fork ?
> >> But the child shall be traced, so debugger would get the SIGTRAP on
> >> the attach on fork returning to usermode. I remember that I explicitely
> >> tested this when adding followfork changes.
> >
> > Yes, the debugger gets SIGTRAPs. The problem arises when the real 
> > parent of the forked process has the code to collect termination 
> > status. Since attaching to a process changes the parent/child 
> > relationships, we need to keep track of the children lost due to 
> > re-parenting so we can properly attribute their exit status to the 
> > "real" parent.
> >
> I recall that someone brought a topic in the list said that this should 
> be fixed, debugging a process should not change
> parent-child relation, instead a new link list data structure should be 
> added to struct proc  to trace debugged process,
> this will make code clean with a small memory overhead.

Yes, I have some old patches to start on this, but I hadn't really finished
them, and it makes wait() a bit more complicated.  It would be nice if
ptrace() had its own pwait() or some such instead of overloading wait().

-- 
John Baldwin



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?201202071340.01126.jhb>